/Projects

Executions & Drafts.

Live builds, planned builds, and the occasional scrapped one. I keep it honest.

Live

Total: 02

[01]● IN PROGRESS

Python-based ML model forecasting short-term price direction on FTSE 100 tickers using volatility and momentum features.

PYTHONPYTORCHPANDASFINANCE

/ Problem

Quant tooling is locked behind institutional walls. Can a teenager replicate the basics with open data?

/ Approach

Engineer momentum + volatility features, train a small LSTM, validate against a walk-forward backtest.

/ Stack

Python, PyTorch, pandas, yfinance. Notebook-driven for now.

/ Next

Plug in a news-sentiment feature using a small transformer.

[02]● BETA

AI-driven property valuation tool trained on hundreds of thousands of images and listing records to estimate fair market value.

PYTHONMLCOMPUTER_VISIONREAL_ESTATE

/ Problem

Property valuations are slow, opaque, and inconsistent between agents.

/ Approach

Train a multi-modal model on listing photos + structured metadata to predict valuations directly from inputs.

/ Stack

Python, PyTorch, image embeddings, tabular features.

/ Next

Expand the training set and tighten error bands on edge cases.

Queue

● QUEUED

A clean React dashboard for tracking a personal paper-trading portfolio.

● QUEUED

Daily auto-summarised market news, shipped to my inbox before school.